More curl tests - #23333
Conversation
|
@NickSdot Could you take a look at this? |
NickSdot
left a comment
There was a problem hiding this comment.
Cannot really judge in detail what's tested, but found some things worth pointing out. Also, can confirm that this closes coverage gaps. ✌️
Coverage: ext/curl
Base: c621cbe27ffe59ed98228724d674dcadcf7f6e8d upstream/master
Tree: 2aa80efcacc776c87712a9beed9487c40be6cffc working tree
+--------+-------+---------+--------------------+-------------------+--------+---------+
| | Tests | Sources | Lines | Branches | Time | Memory |
+--------+-------+---------+--------------------+-------------------+--------+---------+
| Base | 189 | 6 | 2520/2912 (86.54%) | 847/4640 (18.25%) | 64.23s | 45.5 MB |
| Tree | 197 | 6 | 2538/2912 (87.16%) | 857/4640 (18.47%) | 68.17s | 45.1 MB |
| Change | +8 | 0 | +18 / -0 (+0.62%) | +10 / -0 (+0.22%) | +3.94s | -0.4 MB |
+--------+-------+---------+--------------------+-------------------+--------+---------+
|
@LamentXU123 Could you take a look at this? |
|
Sorry, I am not familiar with curl API. Maybe @devnexen can help here. |
|
Windows test fails with:
I will look into making this work on Windows as well. |
Allow setting target of redirect in test
It has the option, but it works differently. We could make it work on earlier versions by including a Content-Type header, but I don't really see the use of that. The goal of the test is to see whether curl_off_t are correctly passed off to curl, not the behavior of curl itself.
They are pretty verbose and don't add much
|
The test FTP server uses fork, so of course it doesn't work on Windows. I added a SKIPIF. After I fixed that, it failed on sapi/cli/tests/php_cli_server_expect_100_continue_socket.phpt. That makes me pretty nervous. I looked into it and cannot reproduce it. I retriggered the CI and now it succeeds. In any case this seems unrelated to these curl tests. So this is done and can be reviewed and merged. |
|
will try to have a look this week end. |
No description provided.